[ad_1]
A study published in Radiology showed that an in-depth learning model was able to predict Alzheimer's disease with a specificity of 82% and a sensitivity of 100 % approximately 6 years before diagnosis, using fluorine-18-fluorodeoxyglucose fluorine-based PET imaging studies. ] "Differences in the configuration of glucose uptake in the brain are very subtle and diffuse", Jae Ho Sohn, MD, Department of Radiology and Imaging Biomedical of the University of California at San Francisco, said in a press release. "People are able to find specific biomarkers of the disease, but metabolic changes are a more global and subtle process."
The researchers examined whether a deep learning algorithm could predict the final diagnosis of Alzheimer's disease in patients who were exposed to PET fluorodeoxyglucose fluoride 18 in the brain.
Investigators collected prospective fluorine-18-fluorodeoxyglucose fluoride-based brain images from the Alzheimer's Neuroimaging Disease database, which included 2,109 imaging studies of 1,002 patients and one retrospective independent test set, including 40 imaging studies of 40 patients. They trained the algorithm thoroughly in depth on 90% of the dataset and then tested it on the remaining 10% and the independent test set. The model was evaluated with sensitivity, specificity and operating characteristic of the receiver.
A model of in-depth learning predicts Alzheimer's disease with a specificity of 82% and a sensitivity of 100% about 6 years before diagnosis, using research studies. PET imaging of the brain by fluorine fluorodeoxyglucose 18,
Source: Shutterstock.com [19659008] The in-depth learning algorithm has learned the metabolic patterns related to Alzheimer's disease, according to the press release. The model obtained a surface under the receiver operating characteristic curve of 0.98 (95% CI, 0.94-1) when tested for its ability to predict the final diagnosis of Alzheimer's disease. in the set of independent tests, reported Sohn and his colleagues. The algorithm achieves a specificity of 82% with a sensitivity of 100% to the detection of the disease on average 75.8 months before the final diagnosis
After comparing the performances of the algorithm with those of the radiological readers, the Researchers also found that the algorithm performed better than readers (57%). sensitivity and specificity of 91%; P <0.05).
"We are very pleased with the performance of the algorithm.It has been able to predict every case that has evolved into Alzheimer's disease," Sohn said in the "If we diagnose the disease of Alzheimer's When all the symptoms have come, the loss of brain volume is so great that it is too late to intervene, and if we can detect it sooner, investigators will have the opportunity to find better ways to slow down or even Stop the Pathological Process. " – Savannah Demko
Disclosure s [19659012]: Sohn Reports UCSF Grants (see # 39, complete study for relevant financial statements of other authors).
[ad_2]
Source link
